Developmental PsychologyChildren’s Developing Judgments About the PhysicalManifestations of PowerBrandon F. Terrizzi, Elizabeth Brey, Kristin Shutts, and Jonathan S. BeierOnline First Publication, December 27, 2018. http://dx.doi.org/10.1037/dev0000657

CITATIONTerrizzi, B. F., Brey, E., Shutts, K., & Beier, J. S. (2018, December 27). Children’s DevelopingJudgments About the Physical Manifestations of Power. Developmental Psychology. Advanceonline publication. http://dx.doi.org/10.1037/dev0000657

Children’s Developing Judgments About the PhysicalManifestations of Power

Brandon F. TerrizziCincinnati Children’s Hospital Medical Center, Cincinnati, Ohio

Elizabeth BreyUniversity of Hawai’i at Ma�noa

Kristin ShuttsUniversity of Wisconsin-Madison

Jonathan S. BeierUniversity of Maryland, College Park

When navigating unfamiliar social environments, it is important to identify who is powerful. Determiningwho has power can be challenging because observers may have limited social information, and becausepeople achieve influence for many reasons. In experiments with 3- to 5-year-old children (n � 192) andadults (n � 32), we investigated the developmental origins and conceptual structure of power judgmentsbased on physical appearance. At 3 years of age, children already associated physical strength withexpansive posture; soon after, expansive postures also supported judgments of normative authority andwere joined by similar judgments about masculine facial structure. By the age of 4, children also matchedhigh- and low-power versions of faces and postures together, indicating that they draw connectionsbetween different aspects of more or less powerful appearance. The complexity and timing of thesechanges highlights limitations in current accounts of the origins of adults’ intuitions about powerfulappearances. This study documents several novel developmental patterns that generate new hypothesesabout the mechanisms that support the emergence of children’s intuitions.

Keywords: face perception, nonverbal behavior, social dominance, social status, physical strength

Supplemental materials: http://dx.doi.org/10.1037/dev0000657.supp

“Power” refers to a person’s potential to control and influenceothers (Emerson, 1962; French, 1956; Magee & Galinsky, 2008).Aligning oneself with powerful people can be advantageous, earn-ing physical protection and social privilege (Van Vugt & Grabo,2015; Van Vugt, Hogan, & Kaiser, 2008). Furthermore, ignoringthe desires of a powerful person can be dangerous—one risksphysical reprisal, social sanction, or other negative consequences

(King, Johnson, & Van Vugt, 2009; Van Vugt, 2006). To ensureadaptive outcomes, people must recognize who is powerful, antic-ipate what powerful people want, and respond strategically to theirbehavior (Anderson & Brown, 2010; Anderson & Kilduff, 2009;Halevy, Chou, & Galinsky, 2011; Magee & Galinsky, 2008; Sa-polsky, 2005).

Navigating real-world power dynamics is far from straightfor-ward (Bierstedt, 1950; Flack & de Waal, 2007; Freeman & Am-bady, 2011). Consider the challenges of one’s first day in aworkplace, on the prison yard, or at a new school. Who holdspower? When entering new social contexts, information thatclearly reveals the relative power of specific individuals may besparse. Diagnostic third-party interactions (e.g., seeing one persontell another what to do; Charafeddine et al., 2015; Smith &Hofmann, 2016) may not occur; controlled resources (a privilegeof the powerful; Enright, Gweon, & Sommerville, 2017; vonRueden, Gurven, & Kaplan, 2008, 2011) may be kept out of sight;and conventional signifiers (e.g., special clothing; Gurney,Howlett, Pine, Tracey, & Moggridge, 2017) may require localcultural knowledge to appreciate their meaning. Given theseconstraints, observers should be highly motivated to seek evi-dence about power from information that is more reliablyavailable. One such source of information is a person’s physicalappearance, which can signal the presence of both person-specific capabilities like physical strength and more subjectiveattributes like confidence (Anderson, Brion, Moore, & Ken-

Brandon F. Terrizzi, Division of General and Community Pediatrics,Cincinnati Children’s Hospital Medical Center, Cincinnati, Ohio; ElizabethBrey, Department of Psychology, University of Hawai’i at Ma�noa; KristinShutts, Department of Psychology, University of Wisconsin-Madison; Jon-athan S. Beier, Department of Psychology, University of Maryland, Col-lege Park.

We thank the Port Discovery Children’s Museum in Baltimore, Mary-land for hosting a significant portion of this work. We thank Rachel Kingfor her assistance with preparing stimuli, Jose Limas Rosas for assistancewith collecting data, Shivani Raina for coding the data, and Karlie Finch,Irene Gomez, and Shirley Duong for recruiting participants. Finally, weextend a special thank you to Alexander Todorov at Princeton Universityfor making the face stimuli used in this study publicly available.

Correspondence concerning this article should be addressed to BrandonF. Terrizzi, Division of General and Community Pediatrics, CincinnatiChildren’s Hospital Medical Center, 3333 Burnet Avenue, Cincinnati, OH45229. E-mail: brandon.terrizzi@cchmc.org

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.

Developmental Psychology© 2018 American Psychological Association 2018, Vol. 1, No. 999, 0000012-1649/18/$12.00 http://dx.doi.org/10.1037/dev0000657

1

nedy, 2012; Freeman & Ambady, 2011; Little, Burriss, Jones, &Roberts, 2007; Lukaszewski, Simmons, Anderson, & Roney,2016; Sell, Hone, & Pound, 2012).

An additional challenge of navigating power dynamics is thatpeople derive power from many sources.1 Those with superiorphysical capabilities, such as brute strength, may assert controlthrough physical force or intimidation (Mazur, 1985; Murray &Schmitz, 2011; Sell et al., 2010, 2012). Others may hold pow-erful social roles, such as being a boss, that grant a morenormative basis for their authority (Cheng, Tracy, Foulsham,Kingstone, & Henrich, 2013; Halevy, Chou, Cohen, & Livings-ton, 2012; Henrich & Gil-White, 2001). The relation betweenphysical strength and normative authority is complex: Someindividuals possess one but not the other, while others maydraw power from both sources. Recognizing the basis of aperson’s power and its relevance to a particular context iscritical for making decisions about how to best monitor andrespond to them (Arnott & Elwood, 2009; Chudek, Heller,Birch, & Henrich, 2011; Henrich & Gil-White, 2001; Parker,1974; Pietraszewski & Shaw, 2015; Smith & Price, 1973). Forinstance, during brief, anonymous, private interactions (e.g., alone encounter in a dark alley), there may be little cost toflouting the wishes of an authority figure, unless that figure isalso physically formidable. Likewise, in public, normativelystructured environments (e.g., a boardroom), disobeying anauthority’s commandments may be a poor choice, regardless ofthe authority’s level of personal strength.

Here we examine children’s developing abilities to meet bothof the challenges identified above: determining who is powerfulfrom limited social information like physical appearance andappreciating that power can be based on both physical formi-dability and normative social roles. As reviewed below, muchremains unknown about how children conceive of the interre-lations between different perceptual cues to power and theirvarious interpretations, as well as the developmental mecha-nisms that support the emergence of these links. The aim of thepresent research is to investigate the developmental trajectoryof this conceptual structure, to better understand children’sintuitions about powerful appearances. Doing so will help con-strain theorizing about the basis for our intuitions about thephysical manifestations of power, as such theories must accountfor the conceptual associations and dissociations that may arise.

We focus on children’s developing judgments regarding twoappearance-based cues (i.e., facial structure and body posture)and two specific forms of power (i.e., physical strength andnormative authority). Both facial structure and body posturemeet the criterion for reliable power cues outlined above: Theyare typically available to observers across multiple contexts(e.g., their manifestation does not depend upon the presence ofthird parties, and they can often be viewed from afar). Impor-tantly, as discussed below, both of these cues have been linkedto adults’ power attributions. Similarly, we targeted children’snotions of strength and authority because these characteristicsare most directly related to our working definition of power: thepotential to exert control over others. Although children reasonabout other personal characteristics associated with hierarchy,status, and leadership, not all of these characteristics fit the coreof our definition (e.g., prestige is more closely related toselective trust and epistemic authority; Chudek et al., 2011).

Adults’ Attributions of Power

When adults encounter new people, they infer power frommultiple aspects of appearance. They view people with a moremature, masculine facial structure (e.g., a pronounced brow orjawline) as both physically stronger (Holzleitner & Perrett, 2016;Toscano, Schubert, Dotsch, Falvello, & Todorov, 2016; Toscano,Schubert, & Sell, 2014) and more socially dominant (Lukaszewskiet al., 2016; Oosterhof & Todorov, 2008; Toscano et al., 2016).Adults also recognize that body posture both reflects a person’scurrent feelings of power (Shariff & Tracy, 2009; Tracy, Robins,& Lagattuta, 2005) and that people may communicate their as-sumed status to others through expansive poses (Brey & Shutts,2015; Burgoon, Buller, Hale, & Turck, 1984; Holland, Wolf,Looser, & Cuddy, 2017; Koski, Xie, & Olson, 2015).

Although adults understand that physical strength and authorityare different things, their attributions reveal similar thinking aboutthe physical manifestation of these two forms of power. Forexample, faces that adults describe as socially dominant are similarto those they describe as physically strong (Toscano et al., 2016).Further, although no research has examined whether expansivepostures make bodies appear stronger, there is evidence for linksamong related representations: Adults’ judgments of social dom-inance from expansive posture are mediated by perceptions ofoverall size (Marsh, Yu, Schechter, & Blair, 2009), and sizeestimations can covary with judgments of physical strength(Fessler, Holbrook, & Snyder, 2012; Mattan, Kubota, & Cloutier,2017; Murray & Schmitz, 2011; Yu, Sun, Zhou, Xu, & Shen,2017). Moreover, adults sometimes view people with a muscularphysique as both strong and likely to hold authority and influenceover others (Lukaszewski et al., 2016).

In summary, there is considerable overlap between the conceptsof strength and authority evoked by certain facial structures andbody postures. Moreover, individual appearance-to-power associ-ations are not represented in isolation from one another. Rather,these associations are units within a complex conceptual structurethat draws connections both between and among different aspectsof appearance and different forms of power. Ultimately, develop-mental theories regarding our intuitions about powerful appear-ance must explain how this interwoven conceptual organizationcomes to be.

Developmental Origins of Power Attributions

Already by the preschool years, young children reason aboutpower in a number of ways. They identify the more powerful oftwo characters by using a wide range of cues presented in acteddisplays and verbal descriptions of agents in third-party interac-tions (Bernard et al., 2016; Brey & Shutts, 2015; Charafeddine etal., 2015, 2016; Gülgöz & Gelman, 2017; Over & Carpenter, 2015;

1 Research on this topic often uses the term “dominance” to describe acharacteristic much like how we have defined “power.” Both terms refer tothe ability to control others. However, some research traditions emphasizethat dominance boils down to physical control and intimidation (Henrich &Gil-White, 2001), whereas others maintain that both physical formidabilityand normative authority contribute to an individual’s dominance level(Lukaszewski et al., 2016; Murray, 2014; Murray & Schmitz, 2011).Because we recognize that formidability and authority are independentroutes to control (Cheng et al., 2013), we use the term power here.

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.

2 TERRIZZI, BREY, SHUTTS, AND BEIER

Pietraszewski & Shaw, 2015; Zhao & Kushnir, 2018). Children’sjudgments in these studies affirm that, like adults, they associatepower with interpersonal control. They view powerful charactersas those who prevail in physical contests (Bernard et al., 2016;Pietraszewski & Shaw, 2015) and who establish how others oughtto behave in their presence (Gülgöz & Gelman, 2017; Over &Carpenter, 2015; Zhao & Kushnir, 2018).

Preschool-age children also reason about relative authority andphysical strength in particular. Even 3-year-old children labelsomeone as “in charge” when they have observed or heard de-scriptions of that person telling another person what to do, makingrules, or having decision-making power over others (Bernard et al.,2016; Brey & Shutts, 2015; Charafeddine et al., 2015; Gülgöz &Gelman, 2017). At these same early ages, witnessing two individ-uals’ relative lifting capacities informs children’s predictions ofwho will be able to lift a heavy object in the future (Fusaro,Corriveau, & Harris, 2011) and their personal preferences for acollaborative partner on physical tasks (Hermes, Behne, Bich,Thielert, & Rakoczy, 2018; Hermes, Behne, & Rakoczy, 2015;Hermes et al., 2016). Further, at 3.5 years, children endorse some-one described as physically stronger as the recommended helperfor a task requiring strength (Paulus & Moore, 2011). Thus,although young children’s understanding of strength and authoritylikely continues to develop, their understanding of these distinctsources of power resemble adults’ concepts in important respects.

Recent research has also examined young children’s attributionsof power from different aspects of physical appearance. Mostnotably, preschoolers reason that a person with mature, masculinefacial features is “stronger” than someone who lacks these features(Cogsdill, Todorov, Spelke, & Banaji, 2014) and that a person withbroad, expansive posture is in charge relative to someone exhib-iting diminutive, restrictive posture (Brey & Shutts, 2015;Charafeddine et al., 2015). Additionally, as reported in the sup-plemental material, 3-year-old children also identify people withlarger musculature as stronger than those with smaller muscula-ture. These attributions reveal that, also like adults, children seecorrespondences between certain aspects of physical appearanceand different forms of power.

Although the above research provides an important foundation,it is not yet sufficient for building theories about the developmen-tal processes that give rise to adults’ intuitions about powerfulappearances. Because these studies have focused on children’sattributions from only a single appearance-based cue, or about asingle notion of power, they do not address the full complexity ofthe conceptual structure underlying adults’ judgments. Conse-quently, it is unclear how and when children acquire their initialrepresentations of powerful appearances and what notions ofpower children can or cannot infer from different cues.

The present study was designed to address three outstandingquestions. First, when do children first become sensitive to theindividual aspects of appearance that adults view as powerful?Cogsdill and colleagues (2014) observed sensitivity to power-relevant variation in facial structure in a collection of 3- and4-year-old children, but did not report analyses exploring devel-opmental changes during this period. A more recent study witheven younger children found that 7-month-olds do not visuallyprefer either dominant or subordinate faces (Jessen & Grossmann,2016); however, sensitivity to such appearances may exist withoutpreferring one over the other. This uncertainty about when children

first become sensitive to these aspects of facial structure highlightsthe need to conduct more thorough investigations in early child-hood.

Similar ambiguities arise regarding children’s sensitivity tobody posture. The only study assessing power attributions fromisolated postural information found that children do not use pos-tural expansiveness to judge the relative authority of two peopleuntil 5 to 6 years of age (Brey & Shutts, 2015). Other studiesprovide hints of such abilities at younger ages, but their results areinconclusive. For instance, 4-year-old children and adults labeldisplays combining expansive posture and a smiling expression asprideful (Tracy et al., 2005), and adults also attribute high statusfrom the same displays (Shariff & Tracy, 2009; Tracy & Robins,2004). However, the relative contributions of postural and facialcues to children’s attributions is unknown (Tracy et al., 2005);moreover, there is no evidence that children also link pride withpower, and one study finds that preschoolers view these displays asmore indicative of anger than pride (Nelson & Russell, 2012).Similarly, in another study, 3-year-olds attributed authority in amanner consistent with the relative postural expansiveness ofcharacters, but the stimuli also included other pragmatic cues, suchas one character’s imperative pointing gesture toward the other(Charafeddine et al., 2015). Thus, more research is needed toisolate the conditions under which children younger than 5 aresensitive to power-relevant variation in body posture.

Second, do children draw connections between different aspectsof appearance that provide information about a person’s power?By analogy, infants see correspondences between emotions ex-pressed in faces and postures (Grossmann, 2010; Hock et al.,2017), and facial and postural information jointly contribute toadults’ inferences about a person’s emotional state (de Gelder, deBorst, & Watson, 2015; Hassin, Aviezer, & Bentin, 2013). Thesefindings indicate that infants and adults possess distinct represen-tations of emotional faces and bodies, and that these representa-tions exist within a larger conceptual framework that allows ob-servers to draw connections between them. Similar associationsmay occur for power-related representations, yet no studies haveassessed whether children or adults link powerful faces and bodies.

Third, what range of power concepts do children associate withdifferent aspects of a person’s appearance? Previous investigationsof children’s judgments from visual appearance have only assessedtheir attributions of a single notion of power from a single aspectof appearance (i.e., just strength from facial structure; just author-ity from posture). Given the interrelations among differentappearance-based cues and notions of power in adults’ judgments,children may be developing (or already possess) more integratedintuitions about powerful appearances than previous approacheshave revealed. Knowing the range of meanings that children as-cribe to different cues is an essential step in describing the con-ceptual framework underlying their understanding of powerfulappearances. For instance, if children attribute different types ofpower to faces and postures, it would suggest that distinct devel-opmental mechanisms may underlie the acquisition of face-powerand posture-power associations.

The Present Research

To investigate these issues, we examined both adults’ and chil-dren’s intuitions about how different notions of power (i.e.,

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.

3CHILDREN’S DEVELOPING JUDGMENTS OF POWER

strength and being in charge) describe variation in facial structureand body posture. Experiment 1, conducted with adult participants,served multiple purposes. The main goal was to characterize thefull pattern of adults’ judgments before investigating the samepatterns in children. Although prior work has assessed adults’intuitions about a subset of these cue and attribution combinations,one combination was novel (i.e., physical strength from postures).Additionally, as no research to our knowledge has asked about allcombinations together, the present study illuminates the relativestrengths of these inferences.

Experiments 2 and 3 took two different approaches to assessingchildren’s intuitions about the physical manifestation of power.First, in a task that closely resembled the one that adults com-pleted, we asked children to produce judgments about physicalstrength and normative authority from variation in faces and pos-tures. Second, we designed a task to assess the extent to whichchildren view powerful faces and bodies as “going together.” Weconducted these studies with children from 3 through 5 years ofage, as previous research has revealed developmental changes inchildren’s sensitivities to power-relevant variation in faces andpostures during this period (Brey & Shutts, 2015; Charafeddine etal., 2015; Cogsdill et al., 2014).

Experiment 1

Experiment 1 assessed the separate influences of facial structureand body posture on adults’ attributions of strength and authority.

Method

Participants. Thirty-two adult undergraduate students withnormal or corrected-to-normal vision (16 women; Mage � 20.04years, SD � 1.57 years) participated online for course credit. Fouradditional students who did not complete the study were excluded.

Materials. Participants viewed paired images of two people,side-by-side and facing forward (i.e., not interacting with oneanother). Half of these displays showed two faces differing in their

inherent structure (bodies not visible). The other half showed twobodies differing in their postural expressions (faces not visible).Figure 1 shows example displays and the entire stimulus set isavailable at OSF (Terrizzi & Beier, 2018; https://osf.io/zghyf/).Characters were aligned so that neither seemed taller than theother. All faces occupied a 600 � 400 pixel region of the display.High and low power bodies occupied 730 � 490 and 730 � 270pixel regions, respectively.

Face stimuli were created with FaceGen v3.1 and drawn from alarger database of 175 male, computer-generated faces (Oosterhof& Todorov, 2008). Each face was derived from a computationalmodel capturing the aspects of facial structure that covary withadults’ judgments of a face’s “dominance.” Faces high in domi-nance have mature, masculine features (e.g., pronounced brow andjawline) while faces low in dominance have less mature, lessmasculine features. Adults’ impressions of these faces have beenextensively validated in many studies of dominance perception(Cogsdill et al., 2014; Oosterhof & Todorov, 2008; Toscano et al.,2014, 2016).

Body stimuli were drawn from a larger set of 132 male,computer-generated characters created by the authors using PoserVersion 10. The body pose of each character was manipulatedbased on prior research demonstrating that adults view peopleholding expansive postures as powerful and restrictive postures asless powerful (Brey & Shutts, 2015; Burgoon, Johnson, & Koch,1998; Holland et al., 2017).

We planned to compare participants’ inferences across faces andbodies, as well as to compare attributions of strength and authoritywithin each stimulus type. Thus, we conducted an initial normingstudy to ensure that adults’ perceptions of dominance for the final setof stimuli were matched, both across and within faces and bodies (seeFigure 1 and the supplemental material for details; no adults whoparticipated in the norming study participated in Experiment 1).Figure 2 shows that the final stimulus set had excellent properties.

Procedure and design. The procedures of this and all sub-sequent experiments were approved by the Institutional Review

Figure 1. Examples of each test trial type from the attribution task of Experiments 1 and 2. Face images werecreated by Oosterhof and Todorov (2008). Body images were created by the authors. See the online article forthe color version of this figure.

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.

4 TERRIZZI, BREY, SHUTTS, AND BEIER

Board of The University of Maryland (Titles: Cognitive andSocial Processes in Adults/Children and How Children FormFirst Impressions of Other People; Protocol numbers 357272–12/410239 –11 and 996211–3). On each of 20 trials, participantsindicated whether the more powerful character was on the leftor right. Trials proceeded in four blocks (five trials each). Eachblock presented the same cue (i.e., faces or postures) andrequested the same attribution (i.e., strength or authority) acrosstrials, reflecting a 2 � 2 within-subject design. Before eachblock, we instructed participants to select the person theythought was stronger or in charge, and we provided a cleardefinition of these terms: “A person who is strong is someonewho can lift and move really heavy things” or “A person whois in charge gets to make all the rules and gets to tell others whatto do.”

The initial cue and attribution types were counterbalancedacross participants. Whereas cue type alternated between blocks,the attribution type was the same for the first two blocks and thenswitched. To assign stimuli to blocks, the 10 face and 10 bodypairs were divided into two sets (five face and five body pairseach), whose presentation order was counterbalanced across par-ticipants. Within blocks, the presentation order of images wasdetermined by a random number generator and fixed for all par-ticipants. Across both cues and attributions, the more powerfulcharacter appeared with equal frequency on the left and right.

Coding and scoring. All analyzed participants responded onall trials. These responses were coded for consistency with theanswers suggested by our prior norming study (i.e., selecting theperson with a mature, masculine face or the person exhibiting

broad, expansive posture as the more powerful character). For eachof the four blocks, scores were also calculated reflecting theproportion of trials in which the anticipated response was given.

Statistical approach. For this and all subsequent experi-ments, data analyses were conducted using R version 3.4.4. Foreach experiment, we first conducted preliminary analyses as-sessing whether participants’ responses were influenced byeach study’s counterbalancing factors. In particular, we wereinterested in discovering whether participants’ judgments weremeaningfully influenced by the order in which they saw eithercue type or the order in which they made different attributions.Because these analyses did not reveal any significant maineffects or easily interpretable interactions in any of our exper-iments, these factors were omitted from further analyses.

For the main analysis, we used a Generalized Linear MixedModel (GLMM) fit by maximum likelihood with Laplace approx-imation and a binomial link function to determine the fixed effectsand interactions of Attribution Type and Cue Type on participants’dichotomous (anticipated or unanticipated) judgments on eachtrial. This model was built using the glmer() function of the lme4package (Bates, Maechler, Bolker, & Walker, 2015). This ap-proach accounts for the correlated nature of responses across trials;our model included a random intercept term identifying individualsubjects. Whereas the GLMM analyzed participants’ judgments oneach trial, our follow-up analyses examined the proportions ofparticipants’ “correct” responses across ages and trial types. Toavoid distributional assumptions, these additional planned analy-ses were pursued with two-tailed nonparametric tests: one-sampleWilcoxon’s tests for comparisons to chance (chance � .50) and

Figure 2. Mean dominance and difference scores (High—Low) for the final selection of high and low powerfaces and postures from Experiment 1’s norming study. Error bars are 95% confidence interval.

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.

5CHILDREN’S DEVELOPING JUDGMENTS OF POWER

both paired Wilcoxon’s and unpaired Mann–Whitney tests, asappropriate.

Results

Figure 3 and Table 1 present the proportion of responses in theanticipated direction for each combination of attribution and cuetype. There were main effects of both Attribution Type, �2(1) �12.46, p � .001, and Cue Type, �2(1) � 14.04, p � .001, as wellas a significant interaction between these factors, �2(1) � 36.24,p � .001. Participants attributed both strength and authority (i.e.,being in charge) from facial structure in the predicted manner, butthey only attributed authority from body posture. Their judgmentsabout relative strength from differences in body posture were lessconsistent than judgments about relative strength from facial struc-ture, z � 4.30, p � .001, relative authority from body posture, z �4.06, p � .001, and relative authority from facial structure, z �2.70, p � .007.

Discussion

Our findings converge with those from previous demonstrationsof adults’ attributions from faces. The same mature, masculinefacial features that give rise to impressions of physical strengthalso support impressions of authority (Toscano et al., 2016).Adults’ attributions from faces suggest that they see some connec-tions between these distinct notions of power. However, partici-pants’ impressions of people differing in their outward posturalexpressions also provide novel insight into the limits of these links,

shown here for the first time: Adults saw people with broad,expansive posture as holding more authority, but did not consis-tently view these same people as stronger. The results from Ex-periment 1 provide a baseline for evaluating the development ofchildren’s intuitions about the power of those whom they encoun-ter.

Experiment 2

Experiment 2 investigated developments in 3- to 5-year-oldchildren’s representations of powerful appearances in two differentways. First, children produced judgments about the relative powerof different characters, in a child-friendly version of the task thatadults performed in Experiment 1 (“attribution task”). Next, thesame children completed a task probing whether they thoughtpowerful faces and bodies “go together” (“matching task”).

Previous work has demonstrated that even 3-year-old children(the youngest age we tested) understand the term “in charge” andcan indicate which of two characters holds more authority (Brey &Shutts, 2015; Charafeddine et al., 2015; Gülgöz & Gelman, 2017).Although other research indicates that 3-year-olds reason accu-rately about a person’s physical strength (e.g., Hermes et al.,2016), their ability to deploy this understanding in a format like thepresent attribution task has not been well established. Thus, in aseparate experiment we first confirmed that 3-year-old childrenindicate—with very high consistency—that people with greatermusculature are stronger than people with lesser musculature(reported fully in the supplemental material; see Figure 5 forexample stimuli). This finding indicates that 3-year-olds hold at

Figure 3. Mean (bars) and median (dots) proportion of trials in which participants selected high power facesand bodies for each trial type in Experiment 1’s and Experiment 2’s attribution task. Asterisks indicatesignificant deviations from chance (0.50). Error bars are 95% confidence interval. � p � .05. �� p � .01. ��� p �.001. See the online article for the color version of this figure.

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.

6 TERRIZZI, BREY, SHUTTS, AND BEIER

least one accurate intuition about strong appearances (as muscu-lature and strength actually covary). Moreover, 3-year-olds areclearly capable of attributing strength to individuals on the basis oftheir physical appearance, as is required of them in the presentattribution task.

Method

Participants. Ninety-six 3- to 5-year-old children, dividedequally across three age groups, participated in the study (3-year-olds: Mage � 3.57 years, SD � .31; 4-year-olds: Mage � 4.55years, SD � .27; and 5-year-olds: Mage � 5.50 years, SD � .29).Each age group contained an equal number of boys and girls.Children were typically developing, exposed predominately toEnglish (�70%), and were recruited from a database of families inthe greater Washington, DC area who expressed interest in partic-ipating in developmental studies. Caregivers identified their chil-dren as primarily White (62.1%), Black/African American(15.8%), or Asian (11.6%). The remaining children were identifiedas either Native American, Hispanic, or “Other.” The modal an-nual income of participating families was over $100,000. Childrenparticipated individually in a university research lab and receiveda small toy for attending.

Eight additional children were tested but excluded from analy-ses because they either failed to complete every trial of the attri-bution task (four 3-year-olds and one 4-year-old), or because of atechnical error that disrupted video recording (one 4-year-old andtwo 5-year-olds).

Materials and procedure.Attribution task. All children completed the attribution task

first. Display images were identical to those used in Experiment 1.Each child viewed the displays on a 23” monitor (1,920 � 1,080pixels) at a distance of approximately 55 cm. The monitor was ona child-sized table. The experimenter sat slightly behind and to theleft of the child; the experimenter was, therefore, out of the child’s

line of sight during test trials. The experimenter controlled theexperimental display from a wireless mouse operated in their lap.

The task began with a brief practice session (three trials) toaccustom children to the experimenter’s verbal instructions. Oneach practice trial, children saw two circles (one yellow, onepurple) and were asked to point to the purple circle. All childrenresponded perfectly.

Next, the experimenter explained that they would now viewimages of people. Children were told to look at each personcarefully and then point to the person who was either stronger orin charge. Only one term was used for each block. Children heardthe same definitions of these terms used in Experiment 1. Allchildren confirmed comprehending each definition the first time itwas provided. The experimenter repeated the relevant definitionbefore each block.

On each trial, a fixation cross was visible while the experimenterprovided instructions, “Look at both people carefully, and point tothe person who’s stronger/in charge.” The experiment made surethat the child was looking at the fixation cross before revealing thetest display and did not provide feedback on the child’s selections.

Matching task. Children were encouraged to place images ofindividual faces and bodies together. The face and body imageswere the same stimuli used in the attribution task. The body imageswere presented in a flipbook with a single body on each page.Laminated cut-outs of individual faces could be affixed to thebodies by joining small strips of Velcro, one located on the backof each face and the other above the shoulders of each body. Theexperimenter sat directly across from the child to administer thetask.

The experimenter told the child a brief story—that the newgame had a book with pictures of both faces and bodies but,earlier, all the faces fell off. The experimenter explained that theyneeded the child’s help putting the faces where they belong. Eachtrial began with the flipbook open to two blank pages. Next, the

Table 1Results From Experiments 1 and 2

Attribution task Task correlations

Age group Trial type M SD Mdn z p � df

3-year-olds Faces - strength .56 .50 .60 .66 .224 .28 30Faces - authority .48 .50 .60 .35 .877 .03 30Postures - strength .64 .48 .60 1.67 .030 .11 30Postures - authority .49 .50 .40 .49 .995 .22 30

4-year-olds Faces - strength .84 .36 1.00 5.01 �.001 .54��� 30Faces - authority .69 .46 .80 2.38 .004 .13 30Postures - strength .90 .30 1.00 5.24 �.001 .30 30Postures - authority .72 .45 1.00 2.28 .005 .22 30

5-year-olds Faces - strength .89 .32 1.00 4.71 �.001 .40� 30Faces - authority .81 .40 1.00 3.62 �.001 .61��� 30Postures - strength .91 .28 1.00 5.12 �.001 .25 30Postures - authority .84 .37 1.00 3.81 �.001 .44�� 30

Adults Faces - strength .90 .30 1.00 5.15 �.001Faces - authority .83 .38 1.00 3.94 �.001Postures - strength .56 .50 .60 .49 .325Postures - authority .89 .32 1.00 4.78 �.001

Note. Mean and median proportion of trials on which participants responded in the anticipated manner, on each trial type of the attribution task, followedby one-sample Wilcoxon tests comparing the median response to chance (.50). The rightmost columns display nonparametric Spearman’s correlationsbetween the proportion of anticipated responses on the matching task and each trial type of the attribution task.� p � .05. �� p � .01. ��� p � .001.

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.

7CHILDREN’S DEVELOPING JUDGMENTS OF POWER

experimenter handed the child two face cut-outs, stacked one ontop of the other. For consistency, the more dominant face wasalways on top, and the experimenter prompted the child to look atboth faces carefully. After the child had examined both images, theexperimenter asked the child to place the faces side-by-side on thetable (in whatever order the child wished), so that both could beseen simultaneously. The experimenter then turned the page toreveal two headless bodies, one on each opposing page, andprompted the child to place the images “where they belong.” Theexperimenter did not provide feedback on children’s choices anddid not make any suggestions for how faces and bodies could bealigned.

Design.Attribution task. For each child age group, the experimental

design was identical to the design used in Experiment 1. Childrenparticipated in four 5-trial blocks, with each block focused on asingle Attribution (i.e., strength, in charge) and Cue (i.e., face,body) type.

Matching task. The matching task consisted of 10 trials. Oneach trial, children saw the high and low power versions of a singleface and body pair. The presentation order of the face and bodypairs on each trial was determined by a random number generator andfixed for all participants. The body with expansive posture appearedwith equal frequency on the left and right, in an ABBABBAABAfashion. Children were randomly assigned to one of two conditionsthat varied whether A referred to the left or right.

Coding and scoring.Attribution task. On each trial, children responded by pointing

to the character they thought was more powerful. As with Exper-iment 1, children’s responses were coded for consistency with thedirection expected by the “dominance” ratings of Experiment 1’snorming study. For each of the four blocks, scores were calculatedreflecting the proportion of trials in which the anticipated responsewas given.

Matching task. On each trial, children affixed two cut-outimages of faces to two images of bodies in a flipbook. Children’sresponses were coded from a video recording of the flipbook takenafter the study was concluded. Choices on each trial were codedfor consistency with the direction expected by the dominanceratings of Experiment 1’s norming study. For each trial, scoresreflected the proportion of trials in which the anticipated responsewas given.

Two independent coders (the first author and a research assistantuninvolved in data collection) recorded the responses of all par-ticipants on each trial of both tasks from video recordings of eachsession. Their initial agreement was almost perfect (Attribution� � .98; Matching � � .98), and all discrepancies were resolvedthrough discussion.

Statistical approach. Our approach to analyzing children’s re-sponses was similar to the approach we used to analyze adults’responses in Experiment 1. In the attribution task, we assessed re-sponses using a GLMM as before, with the addition of a fixed effectof Age Group factor (and interactions between it and Attribution andCue Type, as well as their three-way interaction). In the matchingtask, we assessed the effect of Age Group on participants’ dichoto-mous (anticipated or unanticipated) matching decisions across trials.Because both models now featured Age Group as a between-subjectsfixed effect, follow-up analyses also included Bonferroni-correctedDunn’s tests for multiple comparisons across age groups.

Results

Attribution task. Figure 3 presents the proportion of responsesin the anticipated direction for each combination of Attribution andCue Type for each age group. In our main analysis, there were maineffects of Attribution Type, �2(1) � 44.15, p � .001, Cue Type,�2(1) � 4.68, p � .031, and Age Group, �2(2) � 31.87, p � .001, anda significant interaction between Attribution Type and Age Group,�2(2) � 7.33, p � .026. Children selected the anticipated charactermost consistently when attributing strength and when making judg-ments from variation in body posture. Overall, 3-year-olds were lessconsistent in their judgments than 4- and 5-year-olds, zs � 3.82 and5.24, respectively, ps � .001, but 4- and 5-year-olds did not signifi-cantly differ from each other, z � 1.42, p � .235.

To further understand the complete pattern of children’s judgments,and to pursue the interaction of Age Group and Attribution Type, weassessed the responses of children in each age group using a GLMMto determine the fixed effects and interactions of Attribution Type andCue Type on participants’ dichotomous judgments on each trial.Children had stronger intuitions about strength than authority at eachage, but the effect was largest at 4 years: 3-year-olds: �2(1) � 9.62,p � .001; 4-year-olds: �2(1) � 32.17, p � .001; 5-year-olds: �2(1) �10.01, p � .001. Analyzing each age group separately, there were nosignificant effects of Cue Type (all ps � .14) nor interaction betweenfactors (all ps � .36).

Table 1 shows planned nonparametric comparisons of children’santicipated responses on each trial type to chance (chance � 0.50).Three-year-old children did not often respond in the predicted man-ner. There was only one trial type on which they performed abovechance: attributing strength based on posture. In contrast, both 4- and5-year-old children responded above chance on all four trial types,attributing both strength and authority to more masculine faces andexpansive postures. The supplemental material contains planned anal-yses comparing responses on each trial type across ages.

Figure S2 in the supplemental material indicates there was nocluster of children at any age whose most prominent response wasto select less masculine faces or more restrictive postures as eitherstrong or in charge. This indicates that children do not typicallypass through a developmental stage at which they are sensitive topower-relevant variation in facial structure or body posture butmap that variation to appearance in a direction opposite their laterintuitions.

At the suggestion of an anonymous reviewer, we also examinedhow our results varied across the duration of the experiment. FigureS3 in the supplemental material presents scores for each trial type ateach age, grouped according to the block in which children viewedthat trial type. Because of the exploratory nature of these analyses,their small sample sizes, and many uncorrected tests, caution iswarranted in interpreting these patterns. However, it does appear that3-year-olds attributed strength to dominant faces when this trial typeappeared in either the first or second block. As a comparison, 3-year-old children attributed strength from expansive posture in the firstthree (of four total) blocks. Thus, this exploratory view of the datasupports our main conclusion that children have stronger intuitionsabout power-relevant variation in postures compares to facial struc-ture, if not the claim that children are sensitive to postures first.

Matching task. Figure 4 displays children’s performance on thematching task. Our main analysis indicated a significant effect of AgeGroup, �2(2) � 34.66, p � .001. Three-year-old children performed

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.

8 TERRIZZI, BREY, SHUTTS, AND BEIER

significantly worse than both 4- and 5-year-olds, zs � 4.37 and 5.39,respectively, ps � .001, whereas the two older groups did not differfrom each other, z � 1.11, p � .398. Three-year-old children did notmatch powerful faces and bodies at rates exceeding chance, M � .57,SD � .19, Mdn � .55, z � 1.13, p � .104, while both 4- and5-year-old children performed significantly above chance; 4 years:M � .83 SD � .22, Mdn � .90, z � 4.08, p � .001; 5 years: M � .88SD � .22, Mdn � 1.00, z � 4.29, p � .001.

As detailed in Figure S4 in the supplemental material, children atthe youngest ages showed an approximately even distribution ofmatching task scores, suggesting that most did not have a strongintuition about how to complete the task. With increasing age, chil-dren more consistently paired masculine faces with expansive pos-tures. There was no cluster of children at any age who matchedmasculine faces with bodies showing restrictive postures.

Associations between attribution and matching taskperformance. At each age, we examined correlations between theproportion of anticipated responses on each trial type of the attributiontask, and the proportion of anticipated responses on the matching task.Table 1 displays these bivariate Spearman correlations. At 3 years, nocorrelations were significant. At 4 years, there was a positive corre-lation between children’s scores on the matching task and theirattributions of strength from facial structure. At 5 years, children’sscores on the matching task were correlated with their attributions ofstrength from facial structure and their attributions of authority fromboth facial structure and body posture.

Discussion

Experiment 2 provides a number of insights into the develop-ment of children’s reasoning about powerful people and theirappearances. The results from the attribution task suggest that, aschildren age, their attributions of power increasingly incorporatethe appearance-based cues examined here. Moreover, they come toview both masculine facial structure and expansive body postureas evidence for both strength and authority. The results from the

Figure 4. Mean (bars) and median (dots) proportion of trials in which children matched powerful faces andbodies together. Asterisks indicate significant deviations from chance (0.50). Error bars are 95% confidenceinterval. �� p � .01. ��� p � .001. See the online article for the color version of this figure.

Figure 5. Example of stimuli used in the additional experiment reportedin the online supplemental material. Three-year-old children consistentlylabeled the more muscular person as “stronger,” M � .81, SD � .22,Mdn � .80, z � 4.91, p � .001. See the online article for the color versionof this figure.

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.

9CHILDREN’S DEVELOPING JUDGMENTS OF POWER

matching task further suggest that, during this same developmentalperiod, children’s representations of powerful facial structure andbody posture exist within a conceptual framework that allows themto draw connections between different appearance-based cues.Yet, the developmental trajectory documented here is not simplyone of increasingly adult-like intuitions: Unlike adults, even5-year-old children confidently attributed strength from expansive-ness in body posture.

A limitation of Experiment 2 is that children’s participation inthe attribution task may have influenced their performance on thesubsequent matching task. Although the experimenter did notprovide feedback on either, the attribution task may have increasedthe salience of the physical differences between high and lowpower characters or oriented children toward considering strengthor authority when encountering the same stimuli again. Experi-ment 3 addresses these considerations by testing a new group ofparticipants on the matching task alone.

Experiment 3

Experiment 3 was a direct replication of the matching task fromExperiment 2, conducted with a group of children who had notpreviously performed the attribution task. The experimenter madeno comments about power, strength, or authority, and this waschildren’s first encounter with the face and body images. Conse-quently, matching powerful faces and bodies in the predictedmanner could only be based on children’s preexisting expectationsabout how these two aspects of appearance should align.

Method

Ninety-six 3- to 5-year-old children, divided equally acrossthree age groups, participated in the study (3-year-olds, 13 fe-males: Mage � 3.51 years, SD � .29; 4-year-olds, 13 females:Mage � 4.53 years, SD � .31; and 5-year-olds, 23 females: Mage �5.52 years, SD � .27). Children were typically developing, ex-posed predominately to English (�70%), recruited from the samedatabase utilized in Experiment 2 (n � 44) and from a children’smuseum (n � 52) in Baltimore, MD. The demographics of bothsamples were similar. Caregivers identified their children as pri-marily White (67.7%), Black/African American (14.6%), or Asian(7.3%). The remaining children were identified as either NativeAmerican, Hispanic, or Other. The modal annual income forfamilies was over $100,000. Children who participated individu-ally in a university research lab received a small toy. Children whoparticipated individually at the museum did not receive compen-sation.

Eight additional children were tested but excluded from analy-ses because of a video recording error (two 3-year-olds), an ex-perimenter error (one 5-year-old), failing to complete the study(one 4-year-old), parental interference (one 4-year-old), a parent-reported developmental delay (one 4-year-old), or not meeting thelanguage requirements (two 5-year-olds).

Results

Figure 4 displays children’s performance in Experiment 3. Ouranalytic approach was identical to Experiment 2’s matching task.The main analysis did not reveal a significant effect of Age Group,

�2(2) � 4.28, p � .118. However, to provide continuity with ourexamination of the age differences found in Experiment 2, weagain compared responses at each age to chance performance.Both 4- and 5-year-olds matched powerful faces and bodies; 4years: M � .61 SD � .23, Mdn � .60, z � 1.95, p � .013; 5 years:M � .66 SD � .29, Mdn � .70, z � 2.43, p � .003. Three-year-oldchildren did not, M � .54 SD � .18, Mdn � .50, z � .48, p � .330.

Overall, children in Experiment 2, M � .76, SD � .25, Mdn �.90, performed better than children in Experiment 3, M � .61,SD � .24, Mdn � .60; unpaired Mann–Whitney test: z � 4.06, p �.001. Children’s performance did not vary across testing locations(University or Children’s museum; �2(1) � .45, p � .498), norwere there any interactions involving this factor (X2s � .36).

Discussion

Experiment 3 confirmed the findings from Experiment 2. By age4, children matched powerful faces with powerful bodies, demon-strating that they see correspondences between power-relevantvariation in separate aspects of appearance and are capable ofaligning appearance-based cues together. Children’s ability tomatch power-relevant variation in faces and postures does notdepend on prior experience with these specific stimuli or labelingthem in similar ways. However, there was some suggestion thatscores on the matching task in Experiment 2 were influenced bychildren’s recent experience of the attribution task: Without priorexposure to the stimuli, 3-year-old children did not even performmarginally above chance, and older children’s performance inExperiment 3 was lower than in Experiment 2.

General Discussion

We examined children’s and adults’ abilities to meet two basicchallenges of navigating unfamiliar social environments: deter-mining who is powerful from limited social information (i.e.,physical appearance) and recognizing that power can arise frommany sources. Three questions guided our investigation: When dochildren first become sensitive to physical appearance cues thatadults view as powerful? Do children draw connections betweendifferent aspects of appearance that may provide information abouta person’s power? What range of power concepts do childrenassociate with different aspects of appearance? We found that,over development, children view different aspects of appearanceand different notions of power as increasingly connected. How-ever, neither their sensitivities to powerful appearance cues northeir attributions of power develop uniformly. We suggest that thecomplexity of findings documented here is unlikely to be ex-plained by a single developmental mechanism. Instead, our datahelp generate novel, integrative suggestions regarding the originsof our ability to judge and reason about the power of other people.

Detecting Power From Appearance

We used multiple methods to assess children’s developing sen-sitivity to the same physical cues that adults find powerful. First,we examined their attributions of power from variation in bothfacial structure and body posture. Second, we investigated whetherthey associate power-relevant variation in faces with power-relevant variation in postures. This provided a detailed descriptionof how early representations of powerful appearance emerge.

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.

10 TERRIZZI, BREY, SHUTTS, AND BEIER

Children were sensitive to power-relevant aspects of facialstructure by 4 years, choosing faces with more masculine featuresas both stronger and in charge. This finding converges with resultsfrom Cogsdill and colleagues (2014); however, that work pooled3- and 4-year-old children together, potentially obscuring devel-opmental changes during the early preschool years. Because3-year-olds did not make consistent judgments about faces in thepresent study, children may become attuned to power-relevantaspects of facial structure between 3 and 4 years of age.

Sensitivity to power-relevant aspects of body posture emergedearlier, at the youngest age tested: 3-year-olds indicated that peo-ple holding expansive postures were stronger than those with morerestrictive postures. As prior research had only provided clearevidence that 5-year-old children make inferences about powerfrom postural information alone (Brey & Shutts, 2015), this find-ing motivates a substantial update to our understanding of thetimeline along which children first become sensitive to this aspectof powerful appearance.

Claims about the trajectories of children’s emerging sensitivityto powerful faces and postures require confidence that the exper-imental procedures were appropriate for all ages tested and clarityabout what those procedures assessed. Our procedure for investi-gating children’s attributions from isolated physical cues washighly similar to studies in which children as young as 3 yearshave demonstrated consistent power attributions (Brey & Shutts,2015; Charafeddine et al., 2015; Gülgöz & Gelman, 2017). More-over, the supplemental experiment assessing 3-year-olds’ attribu-tions of strength from musculature yielded highly consistent re-sponses, demonstrating that children are not confused by the taskwhen they hold strong intuitions about the associations underinvestigation. Based on these considerations, we believe that thedevelopmental trajectories observed here document real changes inchildren’s explicit reasoning about the physical manifestations ofpower. Yet, we acknowledge it is possible that other proceduresmight reveal sensitivity to powerful appearances at ages youngerthan those reported here. If so, it will be important to determinewhether procedures might also tap different types of representa-tions (e.g., explicit vs. implicit social knowledge).

One benefit of including the matching task in the present studyis that it placed fewer demands on children’s performance than theattribution task. Children were instructed to pair faces and bodiestogether, but their responses did not require communication backto the experimenter. Furthermore, because the experimenter neverindicated what rule should be used for matching, children couldhave responded “correctly” without reflecting explicitly on how toalign the images. Additionally, children may have matched facesand bodies on dimensions other than strength or authority (e.g.,anger or pride; Nelson & Russell, 2012; Tracy et al., 2005). In lightof these considerations, if children were sensitive to power-relevant aspects of facial structure earlier than the attribution taskrevealed, then 3-year-olds might have matched faces and bodies ina consistent manner. The timing of children’s first success—thatis, age 4 in both Experiments 2 and 3—suggests that children firstview power-relevant variation in both facial structure and bodyposture in similar ways at 4 years, but not earlier.

Four-year-old children’s success on the matching task alsocontributes new insights into their reasoning about the physicalmanifestations of power. One functional benefit of seeing corre-spondences between powerful faces and bodies is that an observer

may arrive at a more accurate estimation of someone’s power levelby pooling information from different sources. Our matching taskindicates that children draw connections between power-relevantvariation in distinct perceptual representations, thereby motivatingfuture research investigating what additional computations theseconnections may support. Such research may also reveal whetherchildren represent a direct association between power-relevantvariation in faces and postures, or whether they see correspon-dences because they link each cue with the same notion of power.

Recognizing Multiple Forms of Power

We also examined children’s reasoning about different sourcesof power. Prior research has only demonstrated that young childrenuse variation in facial structure to determine strength and variationin body posture to determine who is in charge (Brey & Shutts,2015; Charafeddine et al., 2015; Cogsdill et al., 2014). We repli-cated these earlier findings, but also tested the complementaryinferences for the first time. Four- and 5-year-olds attributed bothstrength and authority from facial structure and body posture,demonstrating that children attribute multiple notions of powerfrom a single aspect of a person’s physical appearance. Moregenerally, it suggests that preschool-age children’s understandingof power weaves together notions of both physical formidabilityand normative authority.

Several observations suggest that children initially view anindividual’s appearance as more indicative of physical strengththan authority. First, 3-year-olds treated expansive posture asevidence for strength but not authority. Although children mightbetter understand other terms describing authority (e.g., being a“leader”), it is unlikely that a misunderstanding of in charge canaccount for the overall pattern of 3-year-old’s responses. Multiplestudies indicate that 3-year-olds understand the term in charge andthe definition we used to explain its meaning (e.g., Brey & Shutts,2015; Charafeddine et al., 2015; Gülgöz & Gelman, 2017). Sec-ond, even when children inferred both strength and authority fromphysical appearance, they had stronger intuitions about strength.Third, in Experiment 2, of all four cue/attribution combinations,4-year-olds’ ability to match faces and bodies was uniquely asso-ciated with the extent to which they viewed powerful faces asstrong. This is notable because 3-year-olds already infer strengthfrom body posture. When children first detect correspondencesbetween powerful faces and bodies, they may view such appear-ances as characteristically strong, not in charge.

Yet, the judgments made by younger children and adults di-verged from those of 4- and 5-year-old children. This suggests thata fully interwoven conceptual structure, in which both strength andauthority can be inferred from both facial structure and bodyposture, is neither how power understanding first emerges nor howit matures. These observations, as well as those discussed earlier,support new ways of thinking about the origins of our intuitionsabout the physical manifestations of power.

Developmental Mechanisms Producing IntuitionsAbout Powerful Appearances

To date, three general accounts have sought to explain whypeople hold the intuitions about powerful appearances that they do.One position—the evolutionary hypothesis—is that present-day

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.

11CHILDREN’S DEVELOPING JUDGMENTS OF POWER

judgments are based on innate associations that had predictivevalue in the past, even if some are no longer true (Lukaszewski etal., 2016; Sell et al., 2009; Zebrowitz, 2017). This account hasbeen offered to explain both the attribution of leadership qualitiesto particular facial structures (Van Vugt & Grabo, 2015) and theconceptual link between formidability and authority (Lukaszewskiet al., 2016). A second position—the overgeneralization hypothe-sis—allows that observers may innately attribute formidabilityfrom appearance, but proposes that they attribute other traits (e.g.,social dominance) via heuristic processes that lead them to infer“too much from too little” (Todorov, Olivola, Dotsch, & Mende-Siedlecki, 2015). Additionally, because some permanent featuresof appearance resemble momentary emotional expressions (e.g., apronounced brow may look like a flash of anger), observers mayattribute other personal characteristics (e.g., aggressiveness) fromwhich they infer additional forms of power (e.g., social domi-nance). Lastly, a third position—the cultural hypothesis—rejectsevolutionary specializations entirely, suggesting that associationsbetween physical appearance and traits like dominance are culturalconstructions learned during development (Over & Cook, 2018).

There are strengths to each of these accounts. Each identifiesdevelopmental mechanisms that may indeed contribute to ourintuitions about powerful appearances (i.e., evolved biases, gener-alizations from similar representations, and cultural input). More-over, each account is consistent with a curious fact about intuitionsin this domain: Although judgments can be highly consistentacross observers, they are not always accurate (Todorov et al.,2015).

Yet, none of these accounts can explain the full pattern of resultsdocumented here. For example, a limitation of both the overgen-eralization and cultural hypotheses is that they are primarily basedon data regarding impressions from facial information. Conse-quently, they are largely silent on how the development of impres-sions from postures or other body information might relate to thedevelopment of impressions from faces. In principle, these ac-counts could be broadened in scope, as the mechanisms theyidentify might also underlie the emergence of children’s attribu-tions from powerful postures. However, the present data highlightaspects of children’s intuitions that all three accounts would needto explain, such as children’s greater consistency overall whenmaking judgments from postures and about strength.

We suggest that a comprehensive explanation of how children’sintuitions about powerful appearances develop will be moststraightforward if it incorporates the developmental mechanismshypothesized by each of the three accounts summarized above.Furthermore, additional developmental mechanisms not currentlydescribed by these accounts may also be involved, and we believethat our study’s matching task reveals one such possible mecha-nism. We illustrate our proposal in the remainder of this section.

The age at which a capacity first emerges cannot be taken asdecisive evidence for the degree to which it develops from innatefoundations or individual learning. Learning can happen quickly,and innate systems may be timed to coincide with later-occurringbiological or social processes. Nevertheless, an earlier develop-mental trajectory does provide insight into the quantity and qualityof experience that may be necessary for the capacity to arise. Thus,it is informative to consider which of preschoolers’ intuitionsabout powerful appearances may have roots in capacities alreadypresent in infancy—or even in other species.

Children’s initial tendency to infer strength from appearanceaccords well with the view that humans possess early emerging,specialized capacities for assessing the physical formidability ofothers (Sell et al., 2009). However, differences in children’s im-pressions of strength and authority run counter to one of thedistinctive claims of the evolutionary hypothesis, which predictsthat children’s impressions of strength should immediately engen-der impressions of authority (Lukaszewski et al., 2016). Althoughthe present work targeted ages at which children already haveabstract concepts of both strength and authority, their greaterabilities to infer strength from appearance echo the earlier devel-opment of these notions. Demonstrations of infants’ reasoningabout power are best understood as formidability judgments, basedon an agent’s appearance (e.g., body size; Thomsen, Frankenhuis,Ingold-Smith, & Carey, 2011), coalitional strength (Lourenco,Bonny, & Schwartz, 2016; Pun, Birch, & Baron, 2016), or victoryin a physical struggle (Enright et al., 2017; Gazes & Hampton,2015; Mascaro & Csibra, 2012, 2014). Only after their secondbirthdays do children identify a more powerful individual on thebasis of others’ deferential behavior, as opposed to physical dom-ination (Thomas, Thomsen, Lukowski, Abramyan, & Sarnecka,2018). One way to interpret these findings is that children’s earliestnotions of power center on formidability and include associationsbetween appearance (i.e., body size) and formidability. Later,children may develop notions of authority that are distinct fromformidability and not yet associated with specific aspects of ap-pearance.

Children’s sensitivity to body posture may also be more evolu-tionarily constrained than their intuitions about powerful faces.Many nonhuman species use posture to inflate their apparent sizeduring dominance displays (Mazur, 1985; Weisfeld & Beresford,1982), and, as noted above, preverbal infants use body size topredict the victor of physical contests (Thomsen et al., 2011).These findings suggest that a single, evolved capacity may inferformidability from an individual’s size, and that postural expres-sions capitalize on this system across the animal kingdom. Anopen question is whether posture is initially linked to children’simpressions of power by affecting an individual’s apparent size (asappears to be the case in adults), or whether an expectation thatbody information reflects physical power may heighten children’ssensitivity to culturally provided associations between posture andstrength (a less explored proposal).

Why did children’s sensitivity to powerful facial structureemerge later? Unlike studies on sensitivity to body information,there is currently no wider set of findings with infants and otherspecies to support the proposals of the evolutionary and overgen-eralization hypotheses that innate biases shape our sensitivity topowerful faces. Yet, neither is there much evidence that learnedstereotypes about powerful appearance contribute to early faceattributions, as the overgeneralization hypothesis also proposes—and which is the main driver of developmental change describedby the cultural hypothesis. Future research should first replicatechildren’s delayed sensitivity to powerful facial structure, and theninvestigate the extent to which innate biases and individual learn-ing are involved.

It is also unclear how children come to associate appearancewith authority. Although more research is needed, we suspect thatprocesses of generalization are involved. A recent extension to theovergeneralization hypothesis describes how this might work.

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.

12 TERRIZZI, BREY, SHUTTS, AND BEIER

Whereas the overgeneralization hypothesis has traditionally em-phasized how similar appearances may be mistaken for one an-other, the new perspective proposes that trait judgments fromappearance are also shaped by an observer’s expectations forcorrelations among traits (Stolier, Hehman, & Freeman, 2018).This proposal generates a straightforward empirical prediction: Aschildren develop abstract expectations that strong people holdauthority, they should also attribute authority from the same ap-pearances that they view as indicating strength.

Children’s success on the matching task highlights an addi-tional, previously unconsidered, mechanism that may support thedevelopment of intuitions about powerful appearances: The abilityto see correspondences between power-relevant variation in facialstructure and body posture may make additional generalizationspossible. For instance, children’s early sensitivity to power-relevant variation in posture, coupled with cultural depictionsuniting postural expansiveness with particular facial types, mayactually scaffold their initial detection of power-relevant variationin facial structure. As another example, if children believe thatfacial structure and body posture go together because they bothmanifest an individual’s strength, then additional inferential pro-cesses leading children to attribute authority to some facial struc-tures may entail that similar attributions extend to their associatedpostures. These examples demonstrate how a more detailed de-scription of children’s intuitions, across multiple aspects of ap-pearance, can improve accounts of the mechanisms underlying thedevelopment of our intuitions about powerful appearances. More-over, this approach can be extended to better understand howchildren construct increasingly complex representations of thosewith power. Future work should investigate the timeline alongwhich other power-relevant cues, such as height, musculature, orvocal modulation (Feinberg, DeBruine, Jones, & Little, 2008;Fessler et al., 2012; Han et al., 2017; Pietraszewski, Wertz, Bryant,& Wynn, 2017), are also integrated with the representations ex-amined here.

Reasoning about the physical manifestations of power continuesto develop after the preschool years. Five-year-olds and adultshave similar intuitions about faces but diverge sharply in what theyinfer from body posture. This contrast suggests that the develop-mental mechanisms producing our initial intuitions about powerfulappearances may differ from those that underlie the later refine-ment of these intuitions. Whereas the focus during early childhoodmay be on building up a conceptual structure that increasinglydraws together representations of different aspects of appearanceand notions of power, later development may be geared towardpruning back the least likely of these associations. After all, asadults recognize, even people who are not strong can adopt theposture of those who are in charge.

The contrast between children’s and adults’ attributions alsoraises a novel question about adults’ intuitions: If adults arecapable of scaling back some inaccurate intuitions (e.g., a strength-posture link), why do they appear not to scale back others? Asother researchers have discussed extensively, trait attributionsfrom facial structure are often inaccurate, yet here adults viewedcertain faces as holding more authority than others. Perhaps adultsdistinguish between immutable physical characteristics and ex-pressions that can vary across contexts. Future research mightinvestigate whether impressions from inherent features like facial

structure are more difficult to revise (or suppress) than impressionsfrom variable behavior like posture.

Study Limitations

Several considerations constrain the generalizability of our find-ings. First, our stimuli depicted computer-generated faces andbodies. Although this ensured that the stimuli varied only along thedimensions of interest and did not include elements that mightprovoke unrelated attributions, the artificiality of these images mayhave undermined their ecological validity. For example, the ap-pearance manipulations may have produced faces outside therange children’s typical experiences. Future work might examinewhether children respond more strongly to photo-realistic images,even for more subtle differences in appearance.

Second, our stimuli depicted individuals from a limited range ofsocial identities. Targeting intuitions about male faces and bodiesprovided continuity with prior developmental and adult research(Cogsdill et al., 2014; Oosterhof & Todorov, 2008; Todorov,Mende-Siedlecki, & Dotsch, 2013; Todorov et al., 2015). How-ever, observers’ stereotypes about powerful appearances may varyaccording to a target’s sex, age, race, or other social categorymemberships (Stolier et al., 2018). For example, adults are morelikely to conflate notions of strength and authority when assessingmale, relative to female, faces (Toscano et al., 2016). We do notassume that our findings generalize to judgments about peoplewith other social identities.

Third, although we recruited participants from multiple cities inboth private and public settings, all were part of either urban orsuburban mid-Atlantic communities in the United States, and allwere part of WEIRD culture more broadly (Henrich, Heine, &Norenzayan, 2010). It would be important to know whether ourfindings extend to other cultural contexts, particularly as socialcategory stereotypes are culturally variable. Further, intuitionsabout the sources of normative authority may be more culturallyvariable than intuitions about the sources of physical strength.Future research would benefit from exploring development trajec-tories in communities that diverge in theoretically significant waysfrom the one that we observed. For instance, how might childrenrespond if raised in communities that are female-dominant, egal-itarian, or that encourage gender-neutral socialization (Begler,1978; Goettner-Abendroth, 2018; Shutts, Kenward, Falk, Ivegran,& Fawcett, 2017)? Cross-cultural investigations such as thesewould help to disentangle the contribution of evolved and learnedinfluences on children’s judgments.

Lastly, our procedure does not reveal whether children’s re-sponses were guided by attending to the more or less powerfulcharacter paired within a display. Children may have labeled themore masculine face as stronger by either inferring physicalstrength from its structure directly, or by reasoning disjunctivelyfrom the less masculine face. In the matching task, children mayhave aimed to make either the low- or high-power images gotogether. Our conclusions about children’s sensitivity to variationin posture and facial structure are consistent with any of thesepossibilities. Further research could resolve this ambiguity byasking children to compare high or low power images to a “neu-tral” image.

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.

13CHILDREN’S DEVELOPING JUDGMENTS OF POWER

Conclusion

This study is the most detailed description to date of children’sdeveloping intuitions about the physical manifestations of power.Our results provide multiple insights into the emergence andtrajectory of these intuitions, highlighting the virtues and limits ofexisting developmental accounts. The complexity and timing ofthese changes also motivate new hypotheses about the mechanismsunderlying the emergence of children’s intuitions.

References

Anderson, C., Brion, S., Moore, D. A., & Kennedy, J. A. (2012). Astatus-enhancement account of overconfidence. Journal of Personalityand Social Psychology, 103, 718 –735. http://dx.doi.org/10.1037/a0029395

Anderson, C., & Brown, C. E. (2010). The functions and dysfunctions ofhierarchy. Research in Organizational Behavior, 30, 55–89. http://dx.doi.org/10.1016/j.riob.2010.08.002

Anderson, C., & Kilduff, G. J. (2009). The pursuit of status in socialgroups. Current Directions in Psychological Science, 18, 295–298.http://dx.doi.org/10.1111/j.1467-8721.2009.01655.x

Arnott, G., & Elwood, R. W. (2009). Assessment of fighting ability inanimal contests. Animal Behaviour, 77, 991–1004. http://dx.doi.org/10.1016/j.anbehav.2009.02.010

Bates, D., Maechler, M., Bolker, B., & Walker, S. (2015). Fitting linearmixed-effects models using {lme4}. Journal of Statistical Software, 67,1–48. http://dx.doi.org/10.18637/jss.v067.i01

Begler, E. B. (1978). Sex, status, and authority in egalitarian Society.American Anthropologist, 80, 571–588. http://dx.doi.org/10.1525/aa.1978.80.3.02a00030

Bernard, S., Castelain, T., Mercier, H., Kaufmann, L., Van der Henst, J.-B.,& Clément, F. (2016). The boss is always right: Preschoolers endorse thetestimony of a dominant over that of a subordinate. Journal of Experi-mental Child Psychology, 152, 307–317. http://dx.doi.org/10.1016/j.jecp.2016.08.007

Bierstedt, R. (1950). An analysis of social power. American SociologicalReview, 15, 730. http://dx.doi.org/10.2307/2086605

Brey, E., & Shutts, K. (2015). Children use nonverbal cues to makeinferences about social power. Child Development, 86, 276–286. http://dx.doi.org/10.1111/cdev.12334

Burgoon, J. K., Buller, D. B., Hale, J. L., & Turck, M. A. (1984). Relationsmessages associated with nonverbal behaviors. Human CommunicationResearch, 10, 351–378. http://dx.doi.org/10.1111/j.1468-2958.1984.tb00023.x

Burgoon, J. K., Johnson, M. L., & Koch, P. T. (1998). The nature andmeasurement of interpersonal dominance. Communication Monographs,65, 308–335. http://dx.doi.org/10.1080/03637759809376456

Charafeddine, R., Mercier, H., Clément, F., Kaufmann, L., Berchtold, A.,Reboul, A., & Van der Henst, J.-B. (2015). How preschoolers use cuesof dominance to make sense of their social environment. Journal ofCognition and Development, 16, 587–607. http://dx.doi.org/10.1080/15248372.2014.926269

Charafeddine, R., Mercier, H., Clément, F., Kaufmann, L., Reboul, A., &Van der Henst, J.-B. (2016). Children’s allocation of resources in socialdominance situations. Developmental Psychology, 52, 1843–1857.http://dx.doi.org/10.1037/dev0000164

Cheng, J. T., Tracy, J. L., Foulsham, T., Kingstone, A., & Henrich, J.(2013). Two ways to the top: Evidence that dominance and prestige aredistinct yet viable avenues to social rank and influence. Journal ofPersonality and Social Psychology, 104, 103–125. http://dx.doi.org/10.1037/a0030398

Chudek, M., Heller, S., Birch, S., & Henrich, J. (2011). Prestige-biasedcultural learning: Bystander’s differential attention to potential models

influences children’s learning. Evolution and Human Behavior, 33,46–56.

Cogsdill, E. J., Todorov, A. T., Spelke, E. S., & Banaji, M. R. (2014).Inferring character from faces: A developmental study. PsychologicalScience, 25, 1132–1139. http://dx.doi.org/10.1177/0956797614523297

de Gelder, B., de Borst, A. W., & Watson, R. (2015). The perception ofemotion in body expressions. WIREs Cognitive Science, 6, 149–158.http://dx.doi.org/10.1002/wcs.1335

Emerson, R. M. (1962). Power-dependence relations. American Sociolog-ical Review, 27, 31. http://dx.doi.org/10.2307/2089716

Enright, E. A., Gweon, H., & Sommerville, J. A. (2017). ‘To the victor gothe spoils’: Infants expect resources to align with dominance structures.Cognition, 164, 8–21. http://dx.doi.org/10.1016/j.cognition.2017.03.008

Feinberg, D. R., DeBruine, L. M., Jones, B. C., & Little, A. C. (2008).Correlated preferences for men’s facial and vocal masculinity. Evolutionand Human Behavior, 29, 233–241. http://dx.doi.org/10.1016/j.evolhumbehav.2007.12.008

Fessler, D. M. T., Holbrook, C., & Snyder, J. K. (2012). Weapons make theman (larger): Formidability is represented as size and strength in hu-mans. PLoS ONE, 7, e32751–e32759. http://dx.doi.org/10.1371/journal.pone.0032751

Flack, J. C., & de Waal, F. (2007). Context modulates signal meaning inprimate communication. Proceedings of the National Academy of Sci-ences of the United States of America, 104, 1581–1586. http://dx.doi.org/10.1073/pnas.0603565104

Freeman, J. B., & Ambady, N. (2011). A dynamic interactive theory ofperson construal. Psychological Review, 118, 247–279. http://dx.doi.org/10.1037/a0022327

French, J. R., Jr. (1956). A formal theory of social power. PsychologicalReview, 63, 181–194. http://dx.doi.org/10.1037/h0046123

Fusaro, M., Corriveau, K. H., & Harris, P. L. (2011). The good, the strong,and the accurate: Preschoolers’ evaluations of informant attributes. Jour-nal of Experimental Child Psychology, 110, 561–574. http://dx.doi.org/10.1016/j.jecp.2011.06.008

Gazes, R. P., & Hampton, R. R. (2015). Transitive inference of socialdominance by human infants. Developmental Science, 20, e12367.http://dx.doi.org/10.1111/desc.12367

Goettner-Abendroth, H. (2018). Re-thinking “matriarchy” in modern ma-triarchal studies using two examples: The Khasi and the Mosuo. AsianJournal of Women’s Studies, 24, 3–27. http://dx.doi.org/10.1080/12259276.2017.1421293

Grossmann, T. (2010). The development of emotion perception in face andvoice during infancy. Restorative Neurology and Neuroscience, 28,219–236.

Gülgöz, S., & Gelman, S. A. (2017). Who’s the boss? Concepts of socialpower across development. Child Development, 88, 946–963.

Gurney, D. J., Howlett, N., Pine, K., Tracey, M., & Moggridge, R. (2017).Dressing up posture: The interactive effects of posture and clothing oncompetency judgements. British Journal of Psychology, 108, 436–451.

Halevy, N., Chou, E. Y., Cohen, T. R., & Livingston, R. W. (2012). Statusconferral in intergroup social dilemmas: Behavioral antecedents andconsequences of prestige and dominance. Journal of Personality andSocial Psychology, 102, 351–366. http://dx.doi.org/10.1037/a0025515

Halevy, N., Chou, E. Y., & Galinsky, A. D. (2011). A functional model ofhierarchy. Organizational Psychology Review, 1, 32–52. http://dx.doi.org/10.1177/2041386610380991

Han, C., Kandrik, M., Hahn, A. C., Fisher, C. I., Feinberg, D. R., Hol-zleitner, I. J., . . . Jones, B. C. (2017). Interrelationships among men’sthreat potential, facial dominance, and vocal dominance. EvolutionaryPsychology. Advance online publication. http://dx.doi.org/10.1177/1474704917697332

Hassin, R. R., Aviezer, H., & Bentin, S. (2013). Inherently ambiguous:Facial expressions of emotions, in context. Emotion Review, 5, 60–65.http://dx.doi.org/10.1177/1754073912451331

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.

14 TERRIZZI, BREY, SHUTTS, AND BEIER

Henrich, J., & Gil-White, F. J. (2001). The evolution of prestige: Freelyconferred deference as a mechanism for enhancing the benefits ofcultural transmission. Evolution and Human Behavior, 22, 165–196.http://dx.doi.org/10.1016/s1090-5138(00)00071-4

Henrich, J., Heine, S. J., & Norenzayan, A. (2010). Most people are notWEIRD. Nature, 466, 29. http://dx.doi.org/10.1038/466029a

Hermes, J., Behne, T., Bich, A. E., Thielert, C., & Rakoczy, H. (2018).Children’s selective trust decisions: Rational competence and limitingperformance factors. Developmental Science, 21, 1–12. http://dx.doi.org/10.1111/desc.12527

Hermes, J., Behne, T., & Rakoczy, H. (2015). The role of trait reasoningin young children’s selective trust. Developmental Psychology, 51,1574–1587. http://dx.doi.org/10.1037/dev0000042

Hermes, J., Behne, T., Studte, K., Zeyen, A.-M., Gräfenhain, M., &Rakoczy, H. (2016). Selective Cooperation in Early Childhood - How toChoose Models and Partners. PLoS ONE, 11, e0160881. http://dx.doi.org/10.1371/journal.pone.0160881

Hock, A., Oberst, L., Jubran, R., White, H., Heck, A., & Bhatt, R. S.(2017). Integrated emotion processing in infancy: Matching of faces andbodies. Infancy, 22, 608–625. http://dx.doi.org/10.1111/infa.12177

Holland, E., Wolf, E. B., Looser, C., & Cuddy, A. (2017). Visual attentionto powerful postures: People avert their gaze from nonverbal dominancedisplays. Journal of Experimental Social Psychology, 68, 60–67. http://dx.doi.org/10.1016/j.jesp.2016.05.001

Holzleitner, I. J., & Perrett, D. I. (2016). Perception of strength from 3Dfaces is linked to facial cues of physique. Evolution and Human Behav-ior, 37, 217–229. http://dx.doi.org/10.1016/j.evolhumbehav.2015.11.004

Jessen, S., & Grossmann, T. (2016). Neural and behavioral evidence forinfants’ sensitivity to the trustworthiness of faces. Journal of CognitiveNeuroscience, 28, 1728–1736. http://dx.doi.org/10.1162/jocn_a_00999

King, A. J., Johnson, D. D. P., & Van Vugt, M. (2009). The origins andevolution of leadership. Current Biology, 19, R911–R916. http://dx.doi.org/10.1016/j.cub.2009.07.027

Koski, J. E., Xie, H., & Olson, I. R. (2015). Understanding social hierar-chies: The neural and psychological foundations of status perception.Social Neuroscience, 10, 527–550. http://dx.doi.org/10.1080/17470919.2015.1013223

Little, A. C., Burriss, R. P., Jones, B. C., & Roberts, S. C. (2007). Facialappearance affects voting decisions. Evolution and Human Behavior, 28,18–27. http://dx.doi.org/10.1016/j.evolhumbehav.2006.09.002

Lourenco, S. F., Bonny, J. W., & Schwartz, B. L. (2016). Children andadults use physical size and numerical alliances in third-party judgmentsof dominance. Frontiers in Psychology, 6, 2050. http://dx.doi.org/10.3389/fpsyg.2015.02050

Lukaszewski, A. W., Simmons, Z. L., Anderson, C., & Roney, J. R. (2016).The role of physical formidability in human social status allocation.Journal of Personality and Social Psychology, 110, 385–406. http://dx.doi.org/10.1037/pspi0000042

Magee, J. C., & Galinsky, A. D. (2008). Social hierarchy: The self-reinforcing nature of power and status. The Academy of ManagementAnnals, 2, 351–398. http://dx.doi.org/10.5465/19416520802211628

Marsh, A. A., Yu, H. H., Schechter, J. C., & Blair, R. J. R. (2009). Largerthan life: Humans’ nonverbal status cues alter perceived size. PLoSONE, 4, e5707–e5708. http://dx.doi.org/10.1371/journal.pone.0005707

Mascaro, O., & Csibra, G. (2012). Representation of stable social domi-nance relations by human infants. Proceedings of the National Academyof Sciences of the United States of America, 109, 6862–6867. http://dx.doi.org/10.1073/pnas.1113194109

Mascaro, O., & Csibra, G. (2014). Human infants’ learning of socialstructures: The case of dominance hierarchy. Psychological Science, 25,250–255. http://dx.doi.org/10.1177/0956797613500509

Mattan, B. D., Kubota, J. T., & Cloutier, J. (2017). How social statusshapes person perception and evaluation: A social neuroscience perspec-

tive. Perspectives on Psychological Science, 12, 468–507. http://dx.doi.org/10.1177/1745691616677828

Mazur, A. (1985). A biosocial model of status in face-to-face primategroups. Social Forces, 64, 377–402. http://dx.doi.org/10.1093/sf/64.2.377

Murray, G. R. (2014). Evolutionary preferences for physical formidabilityin leaders. Politics and the Life Sciences, 33, 33–53. http://dx.doi.org/10.2990/33_1_33

Murray, G. R., & Schmitz, J. D. (2011). Caveman politics: Evolutionaryleadership preferences and physical stature. Social Science Quarterly,92, 1215–1235. http://dx.doi.org/10.1111/j.1540-6237.2011.00815.x

Nelson, N. L., & Russell, J. A. (2012). Children’s understanding ofnonverbal expressions of pride. Journal of Experimental Child Psychol-ogy, 111, 379–385. http://dx.doi.org/10.1016/j.jecp.2011.09.004

Oosterhof, N. N., & Todorov, A. (2008). The functional basis of faceevaluation. Proceedings of the National Academy of Sciences of theUnited States of America, 105, 11087–11092. http://dx.doi.org/10.1073/pnas.0805664105

Over, H., & Carpenter, M. (2015). Children infer affiliative and statusrelations from watching others imitate. Developmental Science, 18,917–925. http://dx.doi.org/10.1111/desc.12275

Over, H., & Cook, R. (2018). Where do spontaneous first impressions offaces come from? Cognition, 170, 190–200. http://dx.doi.org/10.1016/j.cognition.2017.10.002

Parker, G. A. (1974). Assessment strategy and the evolution of fightingbehaviour. Journal of Theoretical Biology, 47, 223–243. http://dx.doi.org/10.1016/0022-5193(74)90111-8

Paulus, M., & Moore, C. (2011). Whom to ask for help? Children’sdeveloping understanding of other people’s action capabilities. Experi-mental Brain Research, 211, 593– 600. http://dx.doi.org/10.1007/s00221-011-2676-1

Pietraszewski, D., & Shaw, A. (2015). Not by strength alone: Children’sconflict expectations follow the logic of the asymmetric war of attrition.Human Nature, 26, 44 –72. http://dx.doi.org/10.1007/s12110-015-9220-0

Pietraszewski, D., Wertz, A. E., Bryant, G. A., & Wynn, K. (2017).Three-month-old human infants use vocal cues of body size. Proceed-ings of the Royal Society B: Biological Sciences, 284, 20170656. http://dx.doi.org/10.1098/rspb.2017.0656

Pun, A., Birch, S. A. J., & Baron, A. S. (2016). Infants use relativenumerical group size to infer social dominance. Proceedings of theNational Academy of Sciences of the United States of America, 113,2376–2381. http://dx.doi.org/10.1073/pnas.1514879113

Sapolsky, R. M. (2005). The influence of social hierarchy on primatehealth. Science, 308, 648 – 652. http://dx.doi.org/10.1126/science.1106477

Sell, A., Bryant, G. A., Cosmides, L., Tooby, J., Sznycer, D., . . . Gurven,M. (2010). Adaptations in humans for assessing physical strength fromthe voice. Proceedings of the Royal Society B: Biological Sciences, 277,3509–3518. http://dx.doi.org/10.1098/rspb.2010.0769

Sell, A., Cosmides, L., Tooby, J., & Sznycer, D., von Rueden, C., &Gurven, M. (2009). Human adaptations for the visual assessment ofstrength and fighting ability from the body and face. Proceedings of theRoyal Society B: Biological Sciences, 276, 575–584. http://dx.doi.org/10.1098/rspb.2008.1177

Sell, A., Hone, L. S. E., & Pound, N. (2012). The importance of physicalstrength to human males. Human Nature, 23, 30–44. http://dx.doi.org/10.1007/s12110-012-9131-2

Shariff, A. F., & Tracy, J. L. (2009). Knowing who’s boss: Implicitperceptions of status from the nonverbal expression of pride. Emotion, 9,631–639. http://dx.doi.org/10.1037/a0017089

Shutts, K., Kenward, B., Falk, H., Ivegran, A., & Fawcett, C. (2017). Earlypreschool environments and gender: Effects of gender pedagogy in

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.

15CHILDREN’S DEVELOPING JUDGMENTS OF POWER

Sweden. Journal of Experimental Child Psychology, 162, 1–17. http://dx.doi.org/10.1016/j.jecp.2017.04.014

Smith, J. M., & Price, G. R. (1973). The logic of animal conflict. Nature,246, 15–18. http://dx.doi.org/10.1038/246015a0

Smith, P. K., & Hofmann, W. (2016). Power in everyday life. Proceedingsof the National Academy of Sciences of the United States of America,113, 10043–10048. http://dx.doi.org/10.1073/pnas.1604820113

Stolier, R. M., Hehman, E., & Freeman, J. B. (2018). A dynamic structureof social trait space. Trends in Cognitive Sciences, 22, 197–200. http://dx.doi.org/10.1016/j.tics.2017.12.003

Terrizzi, B. F., & Beier, J. S. (2018, September 24). Children’s developingjudgments about the physical manifestations of power. Retrieved fromosf.io/zghyf

Thomas, A. J., Thomsen, L., Lukowski, A., Abramyan, M., & Sarnecka, B.(2018). Toddlers prefer those who win but not when they win by force.Nature: Human Behaviour, 2, 662–669.

Thomsen, L., Frankenhuis, W. E., Ingold-Smith, M., & Carey, S. (2011).Big and mighty: Preverbal infants mentally represent social dominance.Science, 331, 477–480. http://dx.doi.org/10.1126/science.1199198

Todorov, A., Mende-Siedlecki, P., & Dotsch, R. (2013). Social judgmentsfrom faces. Current Opinion in Neurobiology, 23, 373–380. http://dx.doi.org/10.1016/j.conb.2012.12.010

Todorov, A., Olivola, C. Y., Dotsch, R., & Mende-Siedlecki, P. (2015).Social attributions from faces: Determinants, consequences, accuracy,and functional significance. Annual Review of Psychology, 66, 519–545.http://dx.doi.org/10.1146/annurev-psych-113011-143831

Toscano, H., Schubert, T. W., Dotsch, R., Falvello, V., & Todorov, A.(2016). Physical strength as a cue to dominance: A data-driven ap-proach. Personality and Social Psychology Bulletin, 42, 1603–1616.http://dx.doi.org/10.1177/0146167216666266

Toscano, H., Schubert, T. W., & Sell, A. N. (2014). Judgments of domi-nance from the face track physical strength. Evolutionary Psychology,12, 1–18. http://dx.doi.org/10.1177/147470491401200101

Tracy, J. L., & Robins, R. W. (2004). Show your pride: Evidence for adiscrete emotion expression. Psychological Science, 15, 194–197. http://dx.doi.org/10.1111/j.0956-7976.2004.01503008.x

Tracy, J. L., Robins, R. W., & Lagattuta, K. H. (2005). Can childrenrecognize pride? Emotion, 5, 251–257. http://dx.doi.org/10.1037/1528-3542.5.3.251

Van Vugt, M. (2006). Evolutionary origins of leadership and followership.Personality and Social Psychology Review, 10, 354–371. http://dx.doi.org/10.1207/s15327957pspr1004_5

Van Vugt, M., & Grabo, A. E. (2015). The many faces of leadership.Current Directions in Psychological Science, 24, 484–489. http://dx.doi.org/10.1177/0963721415601971

Van Vugt, M., Hogan, R., & Kaiser, R. B. (2008). Leadership, follower-ship, and evolution: Some lessons from the past. American Psychologist,63, 182–196. http://dx.doi.org/10.1037/0003-066X.63.3.182

von Rueden, C., Gurven, M., & Kaplan, H. (2008). The multiple dimen-sions of male social status in an Amazonian society. Evolution andHuman Behavior, 29, 402–415. http://dx.doi.org/10.1016/j.evolhumbe-hav.2008.05.001

von Rueden, C., Gurven, M., & Kaplan, H. (2011). Why do men seekstatus? Fitness payoffs to dominance and prestige. Proceedings of theRoyal Society B: Biological Sciences, 278, 2223–2232.

Weisfeld, G. E., & Beresford, J. M. (1982). Erectness of posture as anindicator of dominance or success in humans. Motivation and Emotion,6, 113–131. http://dx.doi.org/10.1007/BF00992459

Yu, W., Sun, Z., Zhou, J., Xu, C., & Shen, M. (2017). Humans concep-tualize victory and defeat in body Size. Scientific Reports, 7, 44136.http://dx.doi.org/10.1038/srep44136

Zebrowitz, L. A. (2017). First impressions from faces. Current Directionsin Psychological Science, 26, 237–242. http://dx.doi.org/10.1177/0963721416683996

Zhao, X., & Kushnir, T. (2018). Young children consider individualauthority and collective agreement when deciding who can change rules.Journal of Experimental Child Psychology, 165, 101–116. http://dx.doi.org/10.1016/j.jecp.2017.04.004

Received January 31, 2018Revision received September 26, 2018

Accepted October 5, 2018 �

Thi

sdo

cum

ent

isco

pyri

ghte

dby

the

Am

eric

anPs

ycho

logi

cal

Ass

ocia

tion

oron

eof

itsal

lied

publ

ishe

rs.

Thi

sar

ticle

isin

tend

edso

lely

for

the

pers

onal

use

ofth

ein

divi

dual

user

and

isno

tto

bedi

ssem

inat

edbr

oadl

y.

16 TERRIZZI, BREY, SHUTTS, AND BEIER